Problem 899

Summary: Unexpected behavior of DistanceToOut() in solid G4TwistedTubs
Product: Geant4 Reporter: caogf
Component: geometry/solidsAssignee: Vladimir.Grichine
Status: RESOLVED LATER    
Severity: normal    
Priority: P2    
Version: 8.1   
Hardware: PC   
OS: Linux   
URL: http://cern.ch/gcosmo/geant4/bugs/b899.tar.gz

Description caogf 2006-10-22 19:25:08 CEST
We start with the novice example N01. The geometry is simple, a twisted tub
constructed by G4TwistedTubs is placed in the world volume (a box). We use
geantino as the primary particle (only transportation process in the physics
list). If the particle get out of the world volume, print its PreStepPoint and
PostStepPoint information in the UserSteppingAction.
  As a result, if we shoot the geantino at some point (specified in the code)
in the solid G4TwistedTubs, the DistanceToOut calculated by Geant4 is larger
than expected (out of world volume), the step length is infinite, but the
current volume of this track is still in the twisted tub!!! If we use other
real particles with real physics list, our program will crash because of some
secondary particles produced out of the world volume.
  This happens only for some specific points in the twisted tub, for other
points and
most of cases, our program runs normally
  I have a tar file of the modified example N01, but I don't know how to
upload, sorry! If you want, I will send you by email.

Guofu Cao
IHEP, Beijing
Comment 1 Gabriele Cosmo 2006-10-23 02:24:59 CEST
The tar ball is now accessible through the URL specified in this problem report.
Problem to be verified and analysed.
Comment 2 Vladimir.Grichine 2008-02-01 15:09:48 CET
The proposal requires additional analysis concrning the accuracy 
of trapezoid plane flatness.